Monday, January 21, 2019

Neutron Porosity Logging Revisited

A downhole accelerator, updated source-to-detector spacings, and greater detector efficiency are key features of the new IPL Integrated Porosity Lithology tool that enhance the accuracy of formation evaluation, especially neutron porosity measurements.


Sunday, January 13, 2019

Measuring Permeability Anisotropy

Knowing how fluids flow through a reservoir is fundamental to successful management of hydrocarbon reserves. Fluid flow is governed by the permeability distribution. The latest technique for measuring vertical and horizontal permeability uses a multiprobe wireline formation tester. Operated in open hole, this technique provides measurements before the completion is run allowing reservoir management to begin at the earliest stages of a field's development. 

Permeability - the ease with which fluids flow through rock- has long been identified as one of the most important parameters controlling reservoir performance. Yet it is one of the most difficult to measure. If permeability were the same at all places and in all directions- homogenous and isotropic - then measuring the flow through  a sample of rock would reveal its value. However, rock type and grain size may vary through a reservoir leading to variation in permeability. To complicate matters further, measuring permeability parallel to layers of sedimentary rocks may give a different value to a perpendicular measurement. Therefore permeability measured at the same point in the horizontal direction , Kh, may be different from permeability measured in the vertical direction, Kv. This directional dependency on any type of measurement is called anisotropy. A measurement, such as vertical permeability, in the same direction at two distinct points may also be different. Positional dependency is called heterogeneity. Needless to say, in the horizontal plane, horizontal permeability may have a maximum value, Kh, and a minimum value, Kh. Although anisotropy strictly refers to the directional dependency of a measurement, the ratio Kv/Kh is often used to quantify permeability anisotropy.

The anisotropic nature of permeability can affect any process in which a density difference exists between fluids, for example primary production below the bubblepoint, gas cycling, gas or water coning, waterfloods  and many steam process. It can also influence injection and production rates if the anisotropy is severe. Completion and treatment strategies must also take anisotropy into account - for instance placing perforations near oil-water or oil-gas contacts.

The experience of British Gas Exploration & Production Ltd. emphasizes the importance of anisotropy. The company discovered six small satellite fields of the Morecambe gas fields in the Irish Sea. 

The Triassic Sherwood sandstone reservoirs found there are common to all the Morecambe fields and typically 300 ft thick. Underlying this is an extensive aquifer. Most fields have high permeability- horizontally 200 md , but with individual layers up to 18 darcies. Faults close the reservoirs on one or two sides, with dipping beds sealing the remainder.

To predict the rate and direction of water influx into the reservoir, vertical permeability must be measured. The amount of water influx will determine reserves and, therefore, flow rate and revenue. Underestimating reserves will give a lower flow rate and influence project economics. Overestimating reserves will probably involve penalty payments on future gas sales contracts. 

A modeling study by British Gas showed that at high values of anisotropy- high vertical permeability - considerable reserves are trapped behind the rising aquifer. At the other extreme, low anisotropy - low vertical permeability- does not allow recovery of gas from unperforated layers. Optimum recovery occurs when anisotropy is large enough to retard water influx, but still small enough to drain the unperforated layers.

 Perforating policy for these fields will also be determined by anisotropy. If it is high, only the upper reservoir layers will be perforated to avoid water production. But high vertical permeability will allow drainage of unperforated layers. If anisotropy is low, more perforations will be needed to efficiently drain the field. Reperforating wells will probably be expensive as the likely development will use subsea platforms or those not normally manned. Hence the importance of measuring vertical permeability before perforating.

The problem is that anisotropy not only depends on direction, but also may vary with scale. For example, a single crystal may have an atomic structure that is anisotropic to properties such as electric current flow or acoustic propagation. But a piece of rock formed from randomly packed crystals may be isotropic to the same properties measured at a larger scale. At still larger scales, a series of isotropic rock layers, each with different values for these properties, will behave anisotropically. 

The scale dependency of permeability anisotropy is illustrated by measurements taken by British Gas on its South Morecambe gas fields. Permeability measurements of 1-in. [ 2.5-cm] core plugs yield anisotropies of 0.5 to 0.3. However, vertical pressure profiles over a 400-ft [122-m] thick layer in the producing gas reservoir are consistent with anisotropies as small as 0.002.

Such extreme values are caused by layering of rock on a scale smaller than the scale of the measurement - each layer has a different value of permeability, but all contribute to measurement. Two geological features in particular account for this type of anisotropy: crossbedding and shales. 

Crossbedding is the alternate layering of sands of different grain sizes or textures at an acute angle to the major depositional features. There is little difference between the mineral composition of alternating layers.

Shales have small grain size and usually low permeability. Dispersed shale, for example platy illite which blocks pore space, reduces the permeability of most formations, but does not contribute significantly to anisotropy. On the other hand, shale layers reduce or eliminate flow to adjacent formations and therefore contribute significantly to the anisotropy at some scale. 

Anisotropy is also dependent on shale continuity. For example, a continous shale may totally isolate one zone from another, in which case the permeability  anisotropy measured across the  shale will be zero. If, on the otheerr hand, the shale extends only a short distance from the well, the two zones will not be isolated. Fluid will follow a long, tortuous path around the shale, effectively decreasing the permeability measured across it. So the extent of the shale controls the permeability across it.  

Earlier we said that the  ratio Kv/Kh is often used to quantify permeability anisotropy. A more accurate definition would be to call this ratio vertical permeability anisotropy, which is a useful concept for vertical wells where vertical permeability plays such an important role in field development. For horizontal wells, however, the permeability anisotropy in the horizontal planes becomes equally important.  

Horizontal permeability anisotropy is caused by the depositional environment or by fracturees. Where natural fractures are oriented in one direction there will be a significant difference between the horizontal permeability measured , on a reservoir scale, in the direction of the fractures and that measured normal to them. When tectonic stresses are involved, permeability anisotropies may also occur, as microfractures, aligned with the direction of maxium horizontal stress, open up in the direction normal to the stress. It is also believed that stress anisotropy may cause minor permeability anisotropies without the presence of natural fractures by distorting the pore space. 

There are several different methods of obtaining permeability anisotropy, such as core analysis, well testing techniques and wireline formation tester measurements. One well testing technique- vertical interference testing - is successfully used by a wireline formation tester.

Picture above: Vertical interference test. The well is flowed through one set of perforations creating a pressure disturbance in the reservoir. If there is communication across an interval, a monitor pressure gauge at the second set of perforations will respond to the disturbance. The pressure response depends on the vertical permeability and the boundaries of the zone being tested.

In vertical interference testing, a well is flowed at one zone, creating a pressure disturbance through the reservoir. The effects are recorded on pressure gauges some distance away at a second zone in the same well. The pressure response at the second zone depends on several factors: communication between the two zones, vertical and horizontal permeabilities, and reservoir boundaries. Transient analysis of the pressure response reveals horizontal and vertical permeabilities.  

Vertical interference testing was first developed for well testing using two sets of perforations isolated by straddle packers. This method relies on perfect isolation between the intervals being tested- good packer seals and no casing or cement leaks- and is costly if several zones are to be tested. However, the modular design of the MDT Modular Formation Dynamics Tester tool, using various combinations of probes and packers, allow openhole vertical interference tests to be performed faster and at lower cost- although on a smaller scale.

British Gas used the MDT formation tester to perform five vertical interference tests. The tester configuration used a dual-packer module and a single-probe module. The dual-packer module employs two inflatable packers to isolate about 3.3 ft [1 m] of borehole and was used to create the pressure disturbance - the sink pulse. The single-probe module was mounted above to monitor pressure. The effective distance between the sink pulse and monitor probe was 6.5 ft [2 m] . Using a dual-packer module allowed high flow rates with limited pressure drop and also reduced sanding problems as the fluid velocity across the sand face is lower.

Prior modeling, using a range of vertical permeabilities, showed that the largest possible sink pulse would be required to generate a masurable pressure change across the 6.5-ft gap- the only limitation would be possible sand production. A 10,000-cm3 sample chamber was used to generate the sink pulse and high-precision quartz gauge was connected to the monitor probe. The plan was to flow the formation fluid into the sample chamber and monitor pressure at the monitor probe and between the packers. At the end of each test the pump-out module- also used to inflate the packers- could be used to empty the sample chamber.

The interpretation centers on the pressure transient measured at the vertical monitor probe. The amplitude of the pressure pulse originating at the dual-packer module determines the horizontal permeability, and the travel time gives the vertical permeability. 

Results showed that vertical permeability was between one and two orders of magnitude lower than horizontal permeability. Core measurements available at one depth agreed with the vertical interference tests. At another depth, the core data showed a much lower vertical permeability. 

The low permeability layer seen by the core data may not be areally extensive, whereas the pressure response seen by the MDT formation tester sees beyond this - a distance of three to five times the sink to monitor probe spacing is typical- into the more permeable reservoir. This may account for the discrepancy and shows the significant impact the results have on reserves and development options.

Three-probe Test in West Africa

Multiprobe vertical interference tests were conducted for AGIP recherches Congo, West Africa, to measure permeability anisotropy and to identify permeability barriers across reservoir sections.

The tool configuration for the multiprobe formation tester consists of three probes:
  • the sink probe - to induce a pressure pulse in the formation
  • the vertical monitor probe locate 2.3 ft [ 70 cm] above the sink probe and in the same vertical plane.
  • the horizontal monitor probe directly opposite the sink probe.

The monitor probes measure pressure transients induced at the sink probe. AGIP added sample chambers to this setup to recover clean, pressurized samples of formation water.

A typical sequence of events would be to position the tool and set all three probes against the formation. The integrity of each probe packer seal is checked by performing a small-volume drawdown test- a pretest. A good seal for a probe set in a permeable zone is indicated by a pressure response showing a drawdown followed by a buildup formation pressure. Similiar responses at all three probes are required before the interference test is allowed to proceed. The transient pressure data from pretest may be analyzed to obtain local permeability estimates as with previous formation testers. 

It is advantageous - but not necesssary - to have a constant flow rate during an interference test, and this is achieved by the flow control module.  Up to 1000 cm3 of fluid may be withdrawn from the formation at a specified flow rate during a test through either sink probe or the vertical monitor probe- both are connected to the flowline that runs through most MDT tool modules. The flow control module chamber is reset after test, emptying the contents into the borehole- using the pumpout module- or into a large sample chamber. 

Flowing pressure at the probe must be at least 30% of the mud pressure for the flow control module to operate. In some cases, as in depleted or low-permeability formations, the pressure may be too low to sustain a flow rate. An alternative method is to open the sink probe to a sample chamber attached to the tool and estimate the flow rate. One of the AGIP tests was repeated by opening the sink probe directly to a 1-gallon sample chamber, so that the two methods of providing a pressure pulse could be compared.

Interpretation begins as tests are recorded. Communication is indicated by pressure changes at the monitor probes in response to the pressure pulse. The degree of communication is indicated by the magnitude of the pressure drop. The pressure drop at the horizontal and vertical probes provide a quick estimate of anisotropy. 

Values of horizontal and vertical permeabilities come from transient analysis. Transient analysis involves identifying when spherical or radial flow regimes occur, choosing the location of zone boundaries from openhole logs in such a way as to be compatible with the indicated flow regimes, and, finally, estimating reservoir parameters during those flow regimes. 

One method of identifying the flow regimes present employs pressure derivative plots for which a prequisite is the flow rate history. The interpretation of flow regimes then proceeds in a similiar fashion to that during the interpretation of  a well test.

When the flow rate is unknown, an alternative method may be used. It relies on the fact that multiprobe testing measures pressure transients at two distinct locations away from sink. Fluctuations in flow rate will influence the two pressure transient measurements in some related way. The relationship is purely a function of the flow geometry and rock and fluid properties. This relationship- the G-function may be calculated by using both pressure transients. A plot of G-function versus delta time will approach a slope of -1.5 for spherical flow and -1.0 for radial flow. This approach was used to analyze the AGIP job.

Once the flow regimes are identified, specialized plots may be generated for the periods of spherical flow and radial flow. Spherical analysis allows first estimates to be made for horizontal and vertical mobilities and the porosity-compressibility product. Radial analysis give the horizontal mobility-thicknes product.

The initial estimates are used in formation response models coupled to a parameter estimator to arrive at the best estimate of formation parameters and achieve the best match between observed and calculated pressures. The final match is presented as verification plots- pressure versus time and lobe plots (pictures below). For a low plot, the change in pressure at the vertical monitor probe is plotted against the change in pressure at the horizontal monitor probe during both drawdown and buildup. 

 The separation between vertical monitor probe and sink probe - 2 ft [60cm] did not allow AGIP to test across all zones of reduced porosity that were indicated from petrophysical interpretation of the openhole wireline logs. Several vertical interference tests were conducted over the reservoir to evaluate vertical permeability statistically. Although some dry tests were encountered, no permeability barriers were found. 

 Results from the 1-gallon sample chamber test were in good agreement with the flow control test, and were also in good agreement with permeabilities measured by a drillstem test (DST) over this interval.

The anisotropy ratio for one reservoir from core plug data was 0.8 compared to 0.62 using the MDT tool measurements. The MDT tool results were considered to be more representative and have been incorporated by AGIP into their three-dimensional simulation model.

Two-Probe Test in Abu Dhabi

TOTAL used the MDT tool in four wells to measure permeability anisotropy in a Middle East carbonate reservoir prior to a proposed gas injection project. The test were carried out mostly between limestone and dolomite layers where permeability barriers were expected at the lithology change. 

The MDT tool configuration with single-probe modules was used to increase the spacing between the probes to 8 ft, so that each test would cover as much formation as possible. The flow rate source was the pumpout module, which can pump mud filtrate or formation fluids from the reservoir into the borehole.

The results from the drawdown permeabilities compare well to the stonely permeability log recorded by the DSI Dipole Shear Sonic Imager tool and show extreme permeability heterogeneity. However, results from the vertical interference test measurements show significant differences when compared to permeability measurements from cores. The vertical interference test analysis indicates much lower horizontal permeability at the depth at which core data are available.  High core horizontal permeability measurements are most likely caused by vugs and induced fractures and the fact t hat the measurements took place without overburden pressure.

Although core measurements showed vertical permeability to be almost as good as horizontal permeability, scaling up the data did not provide TOTAL with the correct value of anisotropy for their reservoir model- they had to use a much smaller value to match reservoir performance. The MDT tool test results showed reasons for this. Several MDT tests indicated the presence of permeability barriers; other MDT tests indicated that previously suspected barriers were not present. This enabled TOTAL to revise their simulation model for the gas injection program.

A Barrier Removed?

The importance of permeability anisotropy to sound reservoir management is not in dispute. Vertical interference testing with the MDT tool provides measurements of horizontal and vertical permeability early enough to attack problems of well completion design, stimulation planning and horizontal well trajectory. The resolution of the measurement fills the gap between that of well tests and that of core data so that reservoir models may be refined, leading to better field development strategies, such as enhanced oil recovery programs and infill well placement. 



Tuesday, January 8, 2019

The Promise of Elastic Anisotropy chapter 2

In an anisotropic rock, it is debatable whether the fast or slow S-wave velocity should be used - a slow velocity would give a higher closure stress, therefore a higher volume of pumped fluids. The DSI tool indicated about 8% anisotropy in the shale. Amoco engineers designed a fracture job around the fast shear-wave velocity, predicting lower closure stress, and reducing pumped fluid costs from $100,000 to $35,000 per well. Pump-in closure stress test confirmed the lower stress value indicated by the faster S-wave velocity from the DSI tool. Amoco anticipates saving $10,000 to $65,000 per well on the remaining 300 infill wells to be drilled in the field.

 At the slightly larger scale of a few feet to meters, crosswell seismic surveys also sense elastic anisotropy. But while most oilfield experiments employ vertically traveling waves to study elastic properties, crosswell seismic surveys harness horizontally traveling waves. In such a survey at the British Petroleum test site in Devine, Texas, USA, a seismic source was fired in one well to 56 receiver positions in a wwell 100 m [ 330 ft] away.  Then the experiment was repeated at 55 other source positions. Data processing called tomography divided the area between the wells into 56x56 cells and solved for the P-wave velocity in each square, to create a tomogram. Typical tomography, solving for isotropic velocities, reconstructed an image with layer boundaries that correspond to boundaries seen in gamma ray logs. However, allowing the velocities to be anisotropic enhances the results with a clearer tomographic image between wells.

At the Tree and Forest Scales

Most of the experiments designed to capture in-situ elastic properties have been vertical seismic profiles (VSPs) , at the 10-m [33 ft] wavelength scale. Specially planned VSPs reveal elastic anisotropy of both types, TIV and TIH, but mostly fracture-related TIH anisotropy via shear-wave splitting. These studies show a good correlation between fracture azimuth inferred from VSPs and from other measurements, such as borehole imager tools, regional stress data, surface mapping and experiment on cores. Conducting such studies in the marine setting offers a special challenge, because shear waves can not be generated in nor propagate through water. VSPs can, however, record waves that have been converted from P to S by reflection or refraction. Such vertically propagating shear waves then behave predictably by splitting into fast and slow shear waves when they propagate through fractured rock to borehole receivers. 

As desirable as fractures may be for enhancing fluid flow, they are undesirable in caprock shales, where vertical fractures could diminish their integrity as reservoir seals. Geophysicists are looking into ways to identify fractured and unfractured shale caprock, hoping not to see fracture-related anisotropy in them.

More sophisticated walkaway VSPs, called walkaway for short, can measure elastic properties of layer-anisotropic TIV rock in a way that no others can. Most VSPs rely on near-vertical wave propagation. But without nonvertical travel paths, the elastic properties of TIV materials, such as shales, cannot be measured in situ. The walkaway , with its large source-receiver offset and horizontal travel paths, is able to deliver vital information about shale properties.

 A walkaway survey in the South China Sea sampled a compacting shale sequence through more than 180 degree of propagation angles, usually impossible in all but laboratory experiments. The data revealed fine-scale layering induced anisotropy with horizontal P-wave velocities 12% greater than vertical.

The elastic properties of this highly aneliptic anisotropic shale were used to understand the effects of anisotropy on seismic reflection amplitude variation with offset (AVO) analysis. Surface seismic surveys and VSPs typically involve reflections of waves that propagate within 30 degree of vertical. Even in TIV-anisotropic shales, these near-vertical waves would not sense much anisotropy. But in surveys designed to highlight AVO effects, waves often travel at larger reflection angles. Reflection amplitude depends on the angle of reflection , or offset between source and receiver, and the contrast between P- and S-wave velocities on either side of the reflector. In isotropic rocks, some reflectors - especially those where hydrocarbons are involved - have amplitudes that vary with angle of reflection. 

 In anisotropic rocks, there is the additional complication that the P- and S-wave velocities themselves may vary with angle of propagation, again causing AVO. If a propitious AVO signature is encountered , it is vital to know how much is due to hydrocarbon and how much to anisotropy. This dilemma can be resolved by modeling, which simulates the seismic response to a given rock or fluid contrast. Modeling requires knowledge of elastic properties, and correct modeling should include anisotropy. But anisotropy is a scale-dependent effect, and it is best measured at a scale similiar to the VSP or surface seismic experiment being modeled, such as with a VSP. Most  examples of AVO modeling use sonic-scale elastic parameters - sonic log data. But it is possible to envision an anisotropy , especially if it is fracture-related, at a scale larger than the sonic wavelength but smaller than the VSP wavelength. In this case, the anisotropy may be felt by seismic waves but not by sonic waves. 

Another walkaway, by British Petroleum in the North Sea, measured anisotropic properties in a shale overlying a reservoir with an anomalous AVO signature. The elastic properties were used to model the AVO response at the interface between the shale caprock and the oil sand reservoir. The AVO signature seen in the walkaway data fits the anisotropic model. If the caprock had been assumed to be isotropic, a different AVO response to the oil sand would have been seen, and the sand might not have been identified as oil-bearing. The effect of anisotropy on the interpretation of the AVO anomaly had an important bearing on conclusions drawn from a concurrrent study based on 3D surface seismic data in the area. 

Velocity anisotropy is also beginning to find a home in another corner of the surface seismic world, that of processing surveys to obtain images. This process, called migration, requires knowledge of the velocities of the seismic waves to assign a correct spatial position to reflections recorded in time. In the absence of measurements of anisotropic elastic properties, conventional migration schmes include a 5% fudge factor and assume elliptical anisotropy to convert stacking velocities - results from a prior processing step - to migration velocities. A knowledge of velocity anisotropy beyond the 5% fudge factor, essentially knowing the anellipticity , will become more important in turning ray seismics, as seismic waves spend more time in horizontal travel paths.

Harvesting the Forest

These two types of elastic anisotropy, TIV and TIH, impact the oilfield geoscientist as well as the anisotropist. Measurements of layer-induced anisotropic elastic properties are used to refine processing and produce clearer images or to create better models that lead to more accurate interpretation. In the long run, measuring TIV elastic anisotropy improves reservoir description, which in turn promotes efficient hydrocarbon recovery. 

Measuring fracture- and stress-induced TIH anisotropy may have a more direct and far-reaching impact. Just as elastic waves are bound to travel in the direction of maximum stress or open fractures, so are reservoir fluids. The same forces that induce elastic anisotropy give rise to permeability anisotropy. But the tie between these two is not made routinely, nor is it full understood. 

Establishing the elastic-permeability tie for anisotropy requires geophyscisit, reservoir engineers, geologist and petrophysicst to experiment with such a tie, documenting successes and failures. Today, the most basic level of anisotropic description involves only the geophysicist. The description comprises the azimuth of fracture or stress anisotropy, the degree of anisotropy in relative velocity difference, and the velocities of the two shear waves.

At a more sophisticated level the geologist and petrophysicist add the following information to make further links in the rock-fluid tie: lithology from core or logs; age of the reservoir; history of hydrocarbon maturation; azimuth and aperture of fractures seen in image logs; stress direction in the vicinity of the borehole from caliper logs or hydraulic fractures; and the effect of fluid saturation on resistivity anisotropy.

Tuesday, December 11, 2018

The Promise of Elastic Anisotropy

In certain rocks, sound waves travel at different directions. This characteristics, called elastic tence of aligned features such as fractures, microcracks, fine-scale layers or mineral grains. Combining anisotropy from petrophysics, geology and reservoir engineering may reveal a connection between these alignments and paths.

For most of this century, oilfield theory and practice assumed that waves propagate equally fast in all directions. That is, rocks have isotropic wave velocities.  But waves travel through some rock with different velocities in different directions. This phenomenon, called elastic anisotropy, occurs if there is a spatial ordering of crystals, grains, cracks, bedding planes, joints or fractures- essentially an alignment of strengths or weakness - on a scale smaller than the length of the wave. This alignment causes waves to propagate fastest in the stiffest direction. 

The existence of elastic anisotropy has been largely ignored by exploration and production geophysicist - and for good reasons. The effect is often small. With standard surface seismic measurement techniques most reservoir rocks show directional velocity differences of only 3 to 5% , which may often ben neglected. Moreover, processing data under the assumptions of an isotropic earth is already a challenge; the cost of adding the complications of anisotropy must be justified by improvements in the final seismic image. At most, anisotropy has usually been considered noise that must be filtered out, not as a useful indicator of rock properties.

However, with recent advances in acquisition, processing and interpretation of elastic data , the reasons for ignoring anisotropy are no longer valid. New acquisition hardware and measurement techniques designed to highlight anisotropy reveal highly anisotropic velocities in ultrasonic, sonic and seismic data. This article looks at the evidence for anisotropy, the best way to measure it, and how to use it to enhance reservoir description and optimize development. 

The two requirements for anisotropy - alignment in a preferential direction and at a scale smaller that of the measurement - can be understood through anologies. For the effect of alignment, imagine driving a car in an anisotropic city where streets in the north-south direction have a 30-mile-per-hour speed limit, while the east-west streets have a 50-mile-per-hour limit. East-west drivers will spend less time traveling a given distance than north-south drivers. And drivers will take east-west streets whenever possible. In an anisotropic rock, waves do the same thing, traveling faster along layers of cracks than across them. 

For the effect of scale, a less than perfect but interesting analogy is an insect on a leaf in a forest. The insect sees leaves and branches branching off in random direction : up, down, left, right and everywhere in between. A the scale of the insect, there is no preffered direction of tree growth. There are heterogenieties- sharp discontinuities between leaf and no leaf- but at the insect scale the forest is isotropic. However, to an insect a kilometer away from the forest, the trees appear neatly aligned vertically. To it, the anisotropic nature of the forest is revealed.

Similiarly, a small wavelength wave passing through a packet of thick isotropic layers of differing velocities senses the isotropic velocity of each layer.The wave sees discontinuities at each boundary , but if the wave is small enough to fit several wavelengths in every layer, the layers will still appear isotropic, and no alignment of the discontinuities will be apparent. However, a wave with a wavelength much longer than the layer thickness will not sample layers individually, but as a packet. The packet of layers act as an anisotropic material. The orientation of the layer boundaries is now perceived by the wave-and as one of the fastest directions of travel. And if the individual layers are made of aligned anisotropic grains, as is the case with shales, the anisotropic is even more pronounced.

Anisotropy is then one of the few indicators of variations in rock that can even must be studied with wavelengths longer than the scale of the variations. For once, using 100-ft [30-m] wavelength seismic waves, we can examine rock structure down to the particle scale. However, seismic waves are unable to determine whether the anisotropy is due to alignment at the particle scale or at a scale nearer the length of the wave. In the words of one anisotropy specialist, " The seismic wave is a blunt instrument in that it cannot tell us whether anisotropy is from large or small structures."

Two Types of Anisotropy

There are two styles of alignment in earth materials- horizontal and vertical - and they give rise to two types of anisotropy. Two oversimplified but convenient models have been created to describe how elastic properties, such as velocity or stiffness, vary in the two types. In the simplest horizontal, or layered , case, elastic properties may vary vertically , such as from layer to layer , but not horizontally. Such a material is called tranversely isotropic with a vertical axis of symmetry (TIV). Waves generally travel faster horizontally, along layers, than vertically. Detecting and quantifying this type of anisotroy are important for correlation purposes, such as comparing sonic logs in vertical and deviated dwwells, and for borehole and surface seismic imaging and studies of amplitude variation with offset (AVO).

The simplest case of the second type of anisotropy corresponds to a material with aligned vertical weaknesses such as cracks or fractures, or with unequal horizontal stresses. Elastic properties vary in the direction crossing the fractures, but not along the plane of the fracture. Such a material is called tranversely isotropic with a horizontal axis of symmetry (TIH). Waves travelling along the fracture direction- but within the competent rock-generally travel faster than waves crossing the fractures. Identifying and measuring this type of anisotropy yield information about rock stress and fracture density and orientation. These parameters are important for designing hydraulic fracture jobs and for understanding horizontal and vertical permeability anisotropy. 

More complex cases, such as dipping layers, fractured layered rocks or rocks with multiple fracture sets, may be understood in terms of superposition of the effects of t he individual anisotropies. 

Identifying these types of anisotropy requires understanding how waves are  affected by them. Early encounters with elastic anisotropy in rocks were documented about forty years ago in field and laboratory experiments. Many theoretical papers, too numerous to mention, address this subject, and they are not for beginners. However, it's easy to visualize waves propagating in an anisotropic material. First picture the isotropic case of circular ripples that spread across the surface of a pool of water disrupted by the toss of a pebble. In "anisotropic water" , the ripples would no longer be circular, but almost - not quite - an ellipse. Quantifying the anisotropy amounts to describing the shape of the wavefronts with terms such as ellipticity and anellipticity. In anisotropic rocks, waves behave similarly, expanding in nonspherical, not-quite ellipsoidal wavefronts. 

Waves come in three styles , all of which involve tiny motion of particles relative to the undisturbed material: in isotropic media, compressional waves have particle motion parallel to the direction of wave propagation, and two shear waves have particle motion in planes perpendicular to the direction of wave propagation.

In fluids, only compressional waves can propagate, while soilds can sustain both compressional and shear waves.  Compressional waves are sometimes called P waves, sound waves or acoustic waves, and shear waves are sometimes called S waves. The two are recognized as elastic waves. In a given material, compressional waves nearly always travel faster than shear waves.

When waves travel in an anisotropic material, they generally travel fastest when their particle motion is aligned with the material's stiff direction. For P waves, the particle motion direction and the propagation direction are nearly the same. When S waves travel in a given direction in an anisotropic medium, their particle motion becomes polarized in the material's stiff ( or fast) and compliant (or slow) directions. The waves with differently polarized motion arrive at their destination at different times - one corresponding to the fast velocity, one to the slow velocity. This phenomenon is called shear-wave splitting, or shear-wave birefringence - a term, like anisotropy, with origins in optics. Splitting occurs when shear waves travel horizontally through a layered (TIV) medium or vertically through a fractured (TIH) medium.

Since most geophysical applications place the energy source on the surface, waves generally propagate vertically. Such waves are sensitive to TIH anisotropy, and are therefore useful for detecting vertically aligned fractures. Any stress field can also produce TIH anisotropy if the two horizontal stresses are unequal in magnitude. Vertically traveling P waves by themselves cannot detect anisotropy, but by combining information from P waves traveling in more than one direction, either type of anisotropy can be detected. One approach is to combine vertical and horizontal P waves - such as those which arrive at borehole receivers from distant sources. Another technique compares P waves traveling at different azimuths. Two drawbacks to these compressional-wave methods are that horizontal wave propagation is difficult to achieve except in special acquisition geometries, and that travel paths for P waves are different, introducing into interpretation additional potential differences other than anisotropy. Shear waves, on the other hand, allow a differential measurement in one experiment by sampling anisotropic velocities with two polarizations along the same travel path, giving a greater sensitivity for anisotropy than P waves in multiple experiments. 

Compressional and shear waves of all wavelengths can be affected by anisotropic velocities, as long as the scale of the anisotropy is smaller than the wavelength. In the oil field, the scales of measurement parallel those in the analogy of the insect in a tree in a forest- the insect represents the ultrasonic scale, the tree trunk radius is similiar to the sonic scale and the height of the trees is the scale of the borehole seismic wavelength. The following sections describe how anisotropy is being used to investigate rock properties at each of those scales. 

At the Insect Scale

Wavelenghts in most sedimentary rocks are small - 0.25 to 5 mm for 250 kHz ultrasonic laboratory experiments, and they are four times smaller at 1 MHz. Ultrasonic laboratory experiments on cores show evidence for both layering and fracture related anisotropy in different rock types. While shales generally lead the pack in the relative between velocities of a given wave type in fast and slow directions, experimentalists no longer deliver laboratory results in such simple terms. Instead of the two numbers, P- and S-wave velocities, elastic properties are often characterized by plots of velocity variation around some axis of symmetry. This variation of velocity with angle of propagation has implications for the validity of many empirical relationships that have been established, linking velocity to some other rock property. 

Since ultrasonic laboratory measurements at 0.25 to 5 mm wavelength detect anisotropy, this indicates that the spatial scale of the features causing the anisotropy is much smaller than that wavelength. The main cause of elastic anisotropy in shales appears to be layering of clay platelets on the micron scale due to geotropism - turning in the earth's gravity field- and compaction enhances the effect. 

Laboratory experiments also show the effect of directional stresses on ultrasonic velocities, confirming that compressional waves travel faster in the direction of applied stress. One explanation of this may be that all rocks contain some distribution of microcracks, random or otherwise. As stress is  applied, cracks oriented normal to the direction of greatest stress will close, while cracks aligned with the stress direction will open. In most cases, waves travel fastest when their particle motion is aligned in the direction of the opening cracks. 

Measurements made on synthetic cracked rocks show such results. And computer simulations indicate that rock with an initially isotropic distribution of fractures shows anisotropic fluid flow properties when stressed. Fluid flow is greatest in the direction of cracks than remain open under applied stress, but the overall fluid flow can decrease, because cracks perpendicular to the stress direction, which would feed into open cracks, are now closed.

 At the Tree Trunk Scale

Both types of anisotropy, TIV and TIH, are also detected at the next target scale, approximately the size of a borehole radius, with the DSI Dipole Shear Sonic Imager tool.  At this scale, the most common evidence for TIV layering anisotropy comes from different P-wave velocities measured in vertical and highly deviated or horizontal wells in the same formation  -faster horizontally than vertically. But the same can be said for S-wave velocities. For years, whenever discrepancies appeared between sonic velocities logged in vertical and deviated sections, log interpreters sought explanations in tool failure or logging conditions. Now that anisotropy is better understood, the discrepancies can be viewed as additional petrophysics information. Log interpreters expect anisotropy and look for correlation between elastic anisotropy and anisotropy of other log measurements, such as resistivity.

Fracture- or stress-, induced elastic anisotropy has also been detected by sonic logs through shear-wave splitting. In formations with TIH anisotropy, shear waves generated by transmitters on the DSI tool split into fast and slow polarizations. The fast shear waves arrive at the receiver array before the slow shear waves. Also, the amount of shear wave energy arriving at the receivers varies with tool azimuth as the tool moves up the borehole, rotating on its way.

Detecting anisotropy in DSI waveform data is easy, but using the data to compute the orientations of the split shear waves is a bit trickier. If travel time and arrival energy could be measured for every azimuth at every depth, the problem would be solved, but that would require a stationary measurement. Logging at 1800 ft/hour [ 550 m / hr] , the DSI tool fires its shear sonic pulse alternately from two perpendicular transmitters to an array of similiarly oriented receivers, and the pulse splits into two polarizations. As the tool moves up the borehole, four components -from two transmitters to each of two receivers - of the shear wavefield are recorded. The four components measured at every level, along with a sonde orientation from a GPIT General Purpose Inclinometer Tool measurement , can be manipulated to simulate the data that would have been acquired in a stationary measurement. These data determine the fast and slow  directions, but cannot distinguish between the two. Including the travel-time difference information allows identification of the fast shear-wave polarization direction, which in turn is the orientation of aligned cracks, fractures or the maximum horizontal stress. 

In an example from a well operated by Texaco, Inc. in California , the fast shear-wave polarization direction obtained from such DSI measurements corresponds to fracture azimuths extracted from an FMI fullbore formation Microimager image. 

Amoco exploration and Production used information about shear velocities to optimize hydraulic fracture design in the Hugoton field of Kansas, USA. A key parameter for hydraulic fracture design is closure stress. Closure stress is related through rock mechanics models to Poisson's ratio, which is a function of the P- and S-wave velocities.

Thursday, December 6, 2018

Matrix Treatment in Alberta

This case concerns a Suncor Inc. operated gas well, Pine Creek 10-1, in Alberta, Canada. It has a 2493-ft (760-m) horizontal section, drilled through the carbonate reservoir above the water leg to a measured depth of 14,935 ft [4552 m].

Unlike the usual situation, the best porosity of the horizontal section was believed to be at the toe of the well rather than the heel. However, it was also believed that these high-potential zones had been invaded by drilling mud filtrate.To enhance productivity, it was important to ensure that the acid was pumped into the toe of the well to open up fractures and allow the mud to flow out.

To create the required diversion, it was decided to pump a Foam treatment. Foam is pumped into the formation, blocking further entry of the acid and diverting it to unstimulated reservoir. To minimize friction when pumping at the necessary rate, 2 inch coiled tubing was used to deliver the fluids. The relatively large CT diameter also helped avoid lock-up when running into the long horizontal section and offered more pulling potential if the string had become stuck. 

The downhole assembly consisted of a nozzle, two memory gauges separated by a knuckle joint, and a check valve. The knuckle joint added flexibility to an otherwise stiff assembly. Data collected by the gauges were used after the job to analyze the buildup and breakdown of the formation as successive diversion and acid phases were pumped.

A number of factors complicated the choice of acid additives - which is crucial to the success of any matrix treatment. First, as already noted, Suncor suspected that the formation had been invaded by significant quantities of mud filtrate, which contained a strong emulsifier likely to form an emulsion with spent acid. Second, the presence of 25% hydrogen sulfide [ H2S] in the well necessitated the use of corrosion-control additives that may react with other chemicals in the fluid.

 Consequently, extensive compatibility tests were run between the mud and proposed acid systems. The final treatment design included a number of stages:

  • tubing pickle, which is used to clean up the inside of the coiled tubing - 15% hydrochloric acid [HCl] , inhibitor and surfactant. 
  • preflush, to thin the mud in the wellbore - fracturing oil, antisludge agent and nitrogen, creating a foam with a quality of 50%.
  • Mudclean OB solution, to flush out any remaining mud in the well and water-wet the formation  prior to the FoamMAT job - water, surfactant and solvent as a foam of 50% quality.
  • diversion stages - water and surfactant with nitrogen as a 65% quality foam.
  • Squeeze acid - 15% HCl, with inhibitor, surfactant, de-emulsifier, antisludge agent, miscible solvent and H2S scavenger. The total volume of the acid, some 33,025 gal, was determined by a rule of thumb and past experience of a FoamMAT job carried out on a nearby oil well.
  • postjob flush - fracturing oil and nitrogen. Having pickled the CT and negotiated some problems running in hole caused by a hydrate plug, the preflush was pumped with the CT on bottom- at the end of the toe. Once all the preflush had been displaced across the open hole, the well was shut in for about 15 minutes to allow it to soak and then flowed back to recover any mud filtrate. Next the mudclean OB stage was pumped downhole and displaced using nitrogen. The well was then allowed to flow to clean up and another stage was pumped.
 When this had been displaced out of the well, the main treatment commenced. A series of 15 alternating acid - 1585 gal- and diverter 400 gal -stages were pumped at 25 to 80 gal/min. At the same time , the coiled tubing was gradually pulled out of the hole at about 10 ft/ min from the toe to the heel of the well. After pumping a diverter stage, the pumps were shut down for 10 minutes before the next acid was pumped. 

Midway through the job, the well went on a vacuum. To maintain a positive surface pressure and gain maximum information about the treatment, it was necessary to reduce the bottomhole hydrostatic pressure. The foam qualities of the two fluids were adjusted so that the diverter was 70% and the acid 25%. 

Surface pressure was plotted throughout the job to assess the success of the diversion stages. Once all the acid was pumped, the CT was run back to the toe of the well and postjob flush was pumped to break up the foam in the wellbore and hasten the cleanup.

The well was opened up to flow with the gauges still on bottom. During cleanup, the well flowed spent acid and estimated 21,000 gal of mud filtrate. Suncor believes that this mud came out of the natural fractures of the formation. Once the well was cleaned up, the well pressure and temperature were logged using the memory gauges.

The well is currently waiting to be brought into the production, but Suncor estimates that the acid treatment reduced the pressure drop across the reservoir by 435 to 725 psi. By comparing this to pretreatment pressure and rate information, additional gas deliverability due to the treatment is likely to be 2 to 6 million scf/D.